Application And Study On Data Mining In Telecommunication Services

At present, data mining technique is the most powerful method of data analysis in database field. The approach of analysis is to find latent operation rule by setting up mathematical model with known data. In telecommunication service, understanding different customer group's preferences, shopping attitude, price concept is the key point for successful marketing. Therefore, it's very important to obtain customer's business rules by data mining technique in telecommunication business.In this thesis we firstly present the definition and relative concepts of data mining and the applications in the domain of telecommunication business, and give a brief introduction of classical cluster technique and association rule mining technique. Then we give a discussion of some suitable data mining algorithm for customer segmentation and rules abstraction. We also introduce a customer business model based on data mining technique. We adopt classical K-means and FP-growth technique to abstract the rules in telecommunication business. Firstly, we give a clustering analysis according to customer attributes in telecommunication datasets, then divide customers into different types. At last a series rules can be abstracted by association rules mining for further service recommendation.A series of experimental evaluations are given based on this model, and we particularly describe the idea, implementation process and performance evaluation of the algorithm. Experimental results show that it has advantages as using shorter time and getting higher accuracy for large huge datasets.Above all, this thesis fulfilled a project which can apply data mining technique with telecommunication services. We use data mining technique through historical data and mathematical models to the rules mining with the attributes of customer, features of service and payments of customer, so that we can get the common points of customers that subscribe one service and put these rules into the customer recommendation application.